between 8 and 9 A.M. It appears to me, however, that with regard to chronic diseases the observation of one unvarying and limited diurnal period, characterized by a marked increase of mortality, is not the only precise conclusion which can be gathered from the statistical investigation of this question. The whole of the tables show, as Dr. Finlayson has pointed out with regard to his own and Mr. Watson's results, a gradual and persistent rise, culminating in a definite morning climax. Though in no instance is this rise undisturbed by hourly variations, yet it is very well worthy of notice that not only is the morning tendency in all cases distinctly upwards, but the ante-meridian fluctuation in the curve of each table, is decidedly less erratic than the fluctuation which occurs in spaces indicative of post-meridian mortality. It is scarcely necessary to point out how highly probable it is that this determinate upward movement towards a maximum associated with regular fluctuations, is indicative of the influence of some direct agencies producing that tendency and the variations which characterize it, inasmuch as regularly intercurrent modifications in the progress towards a certain vital result, are suggestive of more complex causal influences than those productive of simple tidal changes. In order to represent this comparative condition more vividly, and also to draw attention to a marked depression common to Mr. West Watson's results and those deducible from the West Riding Asylum records, and also approximately expressed in Schneider's conclusions, I shall take the liberty of reproducing from Dr. Finlayson's article his tabulated arrangement of Mr. Watson's figures.

TABLE No. 3.

Hourly distribution of about 13,000 Deaths, summarized by Mr. West Watson and tabulated by Dr. Finlayson.

By again referring to the numerical tabulation of the West Riding Asylum records (Table No. 2), and comparing it with this table, it will be seen that in the former there is a marked decline in the number of deaths between 8 and 9 P.M., and that in the

latter a similar decided fall characterizes the propinquitous hours of 9 to 10 P.M. In Dr. Finlayson's researches this evening decline is not so well marked, but in the comprehensive statistics of Schneider it is of indubitable occurrence. It is impossible to overlook this great diminution in the number of deaths, characterizing these evening hours; especially as it can be clearly shown that the morning rise in mortality and the evening decline, are converse conditions dependent upon converse aspects of the same causation.

The fact that the action of many incalculable influences makes it difficult to determine to an hour the maximum period of death, is illustrated by the regularity of the progression of the numerical curve when the mortality is represented, not for separate hours, but for triads of hours.

TABLE No. 4.

Distribution of 1,680 Deaths in West Riding Asylum over periods

210

230

220

210

200

190

of three hours.

180

180

By spreading the inappreciable influences in this manner over groups of three hours, so as to allow their operation during one hour to counteract their influence during another, the action of primary influences is shown in the production of changes which are characteristically gradual and precise. Thus the table (No. 4) dealing with deaths from chronic diseases in stages of three hours, shows one steady rise towards a maximum reached at 9 A.M., a decline to a minimum from 9 A.M. to noon, maintained through another period of three hours, and followed by a persistent (though limited) rise towards midnight. In the same manner Dr. Finlayson's table of deaths from acute diseases, grouped in periods of three hours, shows a steady and uninterrupted progress through a

maximum period from 3 to 6 A.M., a marked decline from 6 A.M. to noon, a second elevation lasting from 12 to 3 P.M., a gradual fall from 3 to 6 P.M., and a precipitate plunge towards a minimum reached at midnight.

From a general consideration of all the information bearing on hourly changes in the rates of mortality, it appears :-(1) That there are some hours which are associated with a great liability to death. (2) That in acute and chronic diseases the maximum hours of death are widely different. (3) That in chronic diseases a very large proportion of deaths occur at a period which may be said to range through one hour before and one hour after 9 o'clock A.M. (4) That acute diseases are characterized by two daily periods of marked mortality, the first in the dead of night, the second in the afternoon. (5) That diseases grouped without distinction as to the duration of their course, are distinguished by a maximum mortality rather later than that of acute diseases, and an elevated mortality corresponding with the maximum hours of death from chronic diseases.

[The remainder of the paper treats of the subject "in relation to recurrent changes in the activity of vital functions", and is of too distinctly medical a character to be suitable for reproduction here.-ED. J. I. A.]

On the Mortality of Males and Females from Peritonitis at Ages between Fifteen and Fifty. By WILLIAM ROBERTSON, M.D.,

F.R.S.E.

THE want of a national register of still births is one of the causes which in this country interfere with such important inquiries as those that relate to the true mortality due to parturition. No doubt our system of registration of live births is sufficiently perfect; but until a corresponding record of still births is available, it will be impossible to ascertain, with the accuracy necessary for statistical purposes, the number of women who have, within a given period, become mothers. This is the more to be regretted, since the danger to the life of a female is likely to be far greater when she has given birth to a dead child, than when by her labour she has added a unit to the living population of the country.

In registering the deaths from such a disease as peritonitis occurring within a few weeks of a delivery, it is usual, and I think

right, to refer them to "Metria"-a somewhat comprehensive term, including puerperal peritonitis. But when the RegistrarGeneral receives a certificate of the death of a female from peritonitis, and has no information that she has recently been confined, her death cannot of course be referred to metria or childbirth, but must be entered in his Reports under the head of peritonitis. In the course of some recent attempts to estimate the mortality due in Scotland to childbearing, it occurred to me as probable that a very large proportion of the deaths which take place from metria after still-births, and more especially after illegitimate still-births, -must at present figure in our national statistics as deaths from peritonitis; and that, if this surmise were correct, our recorded mortality from peritonitis should, in the case of females aged between 15 and 50, appear to be much heavier than that of males dying of the same disease within the same limits of age.

The use

In order to test this theory, I have drawn up the following tables, in which the mortality of males and of females from peritonitis is separately exhibited for each of the eleven years 1861-1871; first at ages between 15-50; and then at all other ages. For each of the tables, the numbers of the living were computed, as at the middle of each year, by means of logarithmic series, which were easily obtained from the Census numbers of the population at various ages in April 1861 and 1871. of Thomas's Arithmometer made this a very simple matter. The corresponding numbers of the deaths were taken from the RegistrarGeneral's Detailed Annual Reports for each of the eleven years. The mortality per 100,000 living was obtained by the help of Oakes's tables and Thomas's ingenious machine, by "setting up" the reciprocals of the living, and multiplying them by the numbers of the deaths. The last seven figures of the products were of course decimals, and of these two only were preserved. The process just described is one of the short cuts by which the time of the computer is saved, while I think the chance of arithmetical error is at the same time lessened. A cursory glance at these tables will suffice to show that, between the ages of 15 and 50,— i.e., during the whole reproductive period of female life,-the mortality ascribed to peritonitis was, in every one of the eleven successive years, far more serious among women than among men; and that at this period the average annual mortality of females from this cause of death is nearly double that of males. Further, it is seen that, at the earlier and later stages of human life, the mortality from peritonitis seems to be quite unconnected with the

circumstance of sex; that in successive years the greater mortality was sometimes observed among the males, sometimes among the females, and that the average mortality of the whole eleven years was, although the difference was trifling, rather in favour of female life.

So far then it would appear, that statistical evidence corroborates the theory which it was my object to test when I drew up the tables. I cannot, at present, think of any circumstance likely materially to interfere with the natural deduction, namely, that. during the reproductive period of life the deaths of females are, in our registers, more frequently attributed to peritonitis than those of males; and that this cannot be avoided, since, in the absence of any record of recent delivery, when a child has been born dead, the death of its mother, should she succumb to any of the complaints which are grouped together as childbed fever, would, in our death statistics, be necessarily referred to the heading "peritonitis".

There is one disease to which young females are believed to be more prone than young males, and which, in the absence of a post-mortem examination, may not unnaturally be returned in our death schedules as idiopathic peritonitis. But, fortunately, chronic gastric ulcer, or perforating ulcer of the stomach-the insidious and formidable complaint to which I allude-is not often met with; and its comparatively frequent occurrence in young females will not go far in accounting for such differences in the ratios of male and female mortality from peritonitis, as are exhibited in the columns headed (5) of Tables I and II.

I.-MALES, aged from 15 to 50, Mortality of, in Scotland, from
Peritonitis.